Update ocr_engine.py

ocr_engine.py

@@ -4,19 +4,20 @@ import numpy as np
 from PIL import Image
 
 def extract_weight(pil_image: Image.Image) -> str:
-    # Convert PIL to OpenCV image
+    # Convert to OpenCV format
     img = np.array(pil_image.convert("RGB"))
-
-    # Convert to grayscale
     gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
 
-    # Optional: Resize to improve accuracy
+    # Enhance image for OCR
     gray = cv2.resize(gray, None, fx=2, fy=2, interpolation=cv2.INTER_LINEAR)
+    blurred = cv2.GaussianBlur(gray, (5, 5), 0)
+    _, thresh = cv2.threshold(blurred, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU)
 
-    # OCR config for digits only
-    config = "--psm 7 digits"
-    text = pytesseract.image_to_string(gray, config=config)
+    # OCR with config
+    config = "--psm 7 -c tessedit_char_whitelist=0123456789."
+    text = pytesseract.image_to_string(thresh, config=config)
 
-    # Keep digits and decimal
+    # Extract digits and decimal
     weight = ''.join(filter(lambda x: x in '0123456789.', text))
-    return weight if weight else "No valid weight detected"
+
+    return weight.strip() if weight else "No valid weight detected"